> If I have understood the methods correctly, the 3dttest method is only valid for two groups

It's possible but very cumbersome to analyze your data with 3dttest++.

> the 3dANOVA command is valid when there are no covariates.

Correct.

> When I ran 3dMEMA on two groups, the code ran without an error. However, plotting the
> outcome in the AFNI showed no activations.

More accurately, you failed to find strong evidence for activations.

> Here my only two variables are the group and gender which are both categorical variables,
> but I made gender a numerical variable (labeled 1 or 2 for male vs. female).

You should treat both variables as categorical, which they truly are. Let the program handle the two factors properly; otherwise the results could be messed up. Use 3dMVM, not 3dLME, in this case since there is no random-effects component in your case.

Gang



Edited 1 time(s). Last edit at 05/24/2019 10:30AM by Gang.